Quantitative Coronary Angiographic Studies of Patients With Angina Pectoris and Impaired Glucose Tolerance

From April 2000 to June 2002, 1,529 patients were hospitalized due to nonischemic and ischemic heart disease in our facility of the National Cardiovascular Center, a tertiary referral hospital in the northern district of Osaka, Japan. We obtained informed consent and performed quantitative coronary angiography (QCA) in patients with recurrent chest pain associated with electrocardiographic and/or echocardiographic evidence of myocardial ischemia and without contraindications to the administration of iodinated contrast agent (e.g., predialysis state of renal failure). Thus, a total of 914 patients who had organic stenosis were diagnosed with angina pectoris. Of these patients, the 112 patients who previously underwent coronary artery bypass surgery and the 126 patients with chronic total occlusion were excluded because of difficulty of QCA analysis. Among the remaining 676 patients, 144 patients in whom diabetes had previously been diagnosed and who were being treated by diet therapy alone (n = 48), oral hypoglycemic agents (n = 70), or insulin (n = 26) were defined as the treated diabetes group. The average duration of diabetes in this patient group was 17 years. Also, a 75-g oral glucose tolerance test (OGTT) was refused or could not be performed in some very old or severely ill patients who underwent urgent coronary artery bypass surgery, required mechanical circulatory support, or had refractory infectious diseases (n = 142). Finally, 390 patients with angina pectoris underwent a 75-g OGTT and were divided into the following four groups: 117 patients with normal glucose tolerance (NGT), 3 patients with impaired fasting glucose (IFG), 136 patients with IGT, and 134 patients who showed diabetes pattern (preclinical diabetes).

The 75-g OGTT was performed after overnight fasting. Venous blood samples were taken during fasting and at 120 min after the glucose load to measure blood glucose level. The World Health Organization criteria were used for classifying the OGTT results (5) in the present study. The NGT group was defined as having a fasting plasma glucose (FPG) level <6.1 mmol/l and a postprandial glucose level <7.8 mmol/l. The IFG group was defined as having a FPG level ≥6.1 mmol/l but <7.0 mmol/l and a postprandial glucose level <7.8 mmol/l. The IGT group was defined as having a FPG level <7.0 mmol/l and a postprandial glucose level ≥7.8 mmol/l but <11.1 mmol/l. The preclinical diabetes group was defined as having a FPG level ≥7.0 mmol/l and a postprandial glucose level ≥11.1 mmol/l.

Selective coronary angiography was performed in multiple projections after administration of intracoronary nitroglycerin (0.125–0.25 mg). Coronary angiographic measurements were performed using computer-assisted quantitative analysis (CMS-QCA version 4.0; MEDIS medical imaging systems, Leiden, the Netherlands). In addition to the Gensini score, which has been validated previously (6), the following parameters were used for the assessment of morphological characteristics in global coronary trees. We measured the diameters of the middle section in each major coronary artery segment: segments 1, 2, and 3 of the right coronary artery; segments 6, 7, and 8 of the left anterior descending artery; and segments 11 and 13 of the left circumflex artery. For each patient, we calculated the average diameter of these segments, which was defined as averaged vessel diameter (AVD). We also defined segments narrowed to a diameter of ≤1.5 mm as diseased lesions (with the exception of the far distal portion of segments 8 and 13 showing a smooth and regular edge) and determined averaged lesion length (ALL). The data from QCA were assessed by an experienced cardiologist (I.M.), who was blind to the glucose tolerance status. The interobserver correlation coefficient and the percent error were 0.99 and 7.3 ± 4.9% for AVD and 0.93 and 7.4 ± 5.9% for ALL. The intraobserver correlation coefficient and the percent error were 0.93 and 3.6 ± 2.2% for AVD and 0.98 and 3.9 ± 2.6% for ALL.

All the patients underwent history screening, a physical examination and angiographic and laboratory analyses including HbA_1c (A1C), total cholesterol, triglyceride, HDL cholesterol, LDL cholesterol, and creatinine levels. In addition, patients were assessed for the prevalence of coronary risk factors, i.e., hypertension, hyperlipidemia, smoking habits, and family history, and for the presence of diabetes complications, i.e., nephropathy, a history of myocardial and cerebral infarction, and the presence of arteriosclerotic obliterance. BMI was calculated as the weight in kilograms divided by the square of height in meters (2). Blood glucose levels measured 2 h after breakfast or the 75-g OGTT were defined as those in the postprandial state in the present study. Among the four groups (NGT, IGT, preclinical diabetes, and treated diabetes), ANOVA was performed followed by Bonferroni post hoc testing. For univariate analysis, the following clinical variables and risk factors were regarded as covariates: age, sex, glycemic status (fasting, postprandial glucose level, and A1C level), lipid profiles (total cholesterol, triglyceride, HDL cholesterol, and LDL cholesterol levels), creatinine level, BMI, and the use of cardiovascular medications. On the basis of the results of univariate analysis, we performed multivariate logistic regression analysis to investigate the independent predictors of small AVD (<3.0 mm) and long ALL (>20 mm). A P value of <0.05 was considered to indicate statistical significance. All analyses were performed using Stat-View software, version 5.0 (SAS Institute, Cary, NC).

Clinical characteristics are summarized in Table 1. Most of the parameters, with the exception of BMI, triglyceride and HDL cholesterol levels, and the prevalence of β-blocker treatment, were not different between the IGT and NGT groups. The prevalence of renal failure (serum creatinine level ≥177 μmol/l) was higher in the treated diabetes group than in the IGT group.

As shown in Table 1, the fasting glucose and A1C levels were similar between the NGT and IGT groups. However, the postprandial glucose level was significantly different between the NGT and IGT groups (P < 0.05). All of these parameters relating to glucose metabolism were significantly higher in the preclinical and treated diabetes groups than in the IGT group (P < 0.05).

The results of QCA are summarized in Fig. 1. In the IGT group, smaller AVD and longer ALL were found than in the NGT group (P < 0.05). These angiographic changes were more prominent in the treated diabetes group (P < 0.05). When assessed by the Gensini score, statistics reached significance only between the NGT and the treated diabetes groups.

We found that the angiographic characteristics of the IGT and diabetes groups were small AVD and long ALL. Therefore, to investigate which clinical variables and risk factors were associated with these morphological changes in the coronary artery, we performed univariate and multivariate analyses (Table 2). Age, female sex, and postprandial glucose and creatinine levels were significant predictors of small AVD (<3.0 mm) by univariate analysis. By multivariate analysis, female sex and postprandial glucose and creatinine levels were independent predictors. In addition, age, female sex and fasting and postprandial glucose levels, A1C, and lower HDL cholesterol levels were significant predictors of long ALL (>20 mm) by univariate analysis. Among these parameters, age, female sex, and postprandial glucose and lower HDL cholesterol levels were independent predictors by multivariate analysis.

The major findings of the present study using computer-assisted QCA analysis are that the morphological changes of small vessel diameter and diffuse vessel narrowing developed not only in the diabetes groups but also in the IGT group. Furthermore, these changes were associated with postprandial hyperglycemia.

A previous study reported a high prevalence of abnormal glucose tolerance (IGT 35% and diabetes 31%) in 181 patients with acute myocardial infarction (6). In the present study of patients with angina pectoris, the prevalences of IGT and diabetes (preclinical and treated) were 24.9 and 52.5%, respectively. Also, FPG and A1C levels were similar between the IGT and NGT groups. These findings indicate that IGT plays a pathological role and its diagnosis based on a 75-g OGTT is important in patients with coronary artery diseases.

Although IGT is a risk factor not only for the development of diabetes but also for cardiovascular morbidity and mortality (1,7–10), there have been few reports of the angiographic characteristics in patients with IGT. In a previous study of 99 patients with coronary artery disease (11), abnormal glucose tolerance (IGT and diabetes) was found in 37 patients, for whom the degree of coronary atherosclerosis, assessed by a hand-held caliper method, was the same as that for patients with NGT. Another study (12) was performed using 466 patients who were undergoing coronary angiography, including patients with chest pain syndrome (10%) and those with nonischemic cardiac disease (21%). QCA analysis could not reveal differences between the NGT group (n = 291) and the IFG group (n = 82), which would be due to the high percentage of patients without significant coronary artery narrowings (67–73%) in these two groups. Therefore, in the present study, we focused on patients both with angina pectoris and with angiographically documented coronary artery narrowings.

QCA analysis has the advantage of being more accurate and reproducible than visual hand-held caliper measurements (13–16). However, in particular patients with diffuse narrowings it is difficult to identify a reference segment against which diseased lesions can be compared. In these patients, there is the possibility of underestimating the severity of vascular disease when it is assessed by “percent” diameter of stenosis. In fact, as shown in Fig. 1, this could happen with the measurement of Gensini score, which enabled us to detect changes in patients with IGT. In the present study, we defined segments narrowed to a diameter of ≤1.5 mm as diseased lesions in the major proximal segment of coronary trees. Although further modification of methodology to consider left main coronary artery lesions is needed, the present study using QCA analysis, which assesses absolute values, is a novel method to assess pathological changes of diffuse and small coronary lesions.

We found that there were significant differences in AVD and ALL between the NGT and IGT groups, indicating that angiographic atherosclerotic changes (smaller vessel diameter and longer lesion length) develop in coronary arteries in patients with IGT. The degree of changes found in patients with IGT was similar to that in patients with preclinical diabetes but was less than that in treated diabetic patients in whom FPG and A1C levels were markedly elevated.

As seen in previous epidemiological studies, such as the DECODE study (2), the risk of cardiovascular disease is strongly associated with postprandial hyperglycemia. In the present study, multivariate analysis showed that the postprandial glucose level was an independent predictor for both small AVD (<3.0 mm) and long ALL (>20 mm) (17–22), indicating that postprandial hyperglycemia is strongly associated with diffuse and small coronary artery narrowing in patients with abnormal glucose tolerance.

Hyperglycemia promotes atherogenesis by several possible mechanisms, including increased generation of free radicals (oxidative stress), decreased production of nitric oxide, activation of the polyol pathway and the diacylglycerol–protein kinase C system, and increases in nonenzymatic glycation products and the glycosylation of certain proteins (23–28). A recent in vivo study (29) demonstrated that inflammatory cytokine levels were more affected by oscillatory than continuous hyperglycemia and that this inflammatory response to hyperglycemic spikes was attenuated by antioxidants. These findings may explain, in part, why acute hyperglycemia occurring postprandially could produce cardiovascular damage.

Because this is a retrospective clinical investigation, we acknowledge the possibility of a selection bias for some patients. Also, a statistical association does not prove that the risk factor directly promotes coronary artery atherosclerosis. Thus, in addition to measurements of potential confounders (e.g., fat distribution, physical fitness, or inflammation), prospective studies of an interventional nature will be required to determine whether postprandial hyperglycemia has a causal role in development of coronary artery disease.

In summary, the present study using quantitative angiographic analysis demonstrates that postprandial hyperglycemia is associated with the development of diffuse coronary artery narrowings in patients with angina pectoris.